METAMODELS FOR MULTIOBJECTIVE OPTIMIZATION OF VISCOELASTIC DAMPED STRUCTURES

Antonio Marcos Gonçalves de Lima¹; Domingos Alves Rade¹; Noureddine Bouhaddi²

¹ Federal University of Uberlandia; ² FEMTO-ST Institute, UMR

Resumo

The use of viscoelastic materials is an efficient strategy to reduce the vibration levels of mechanical systems. However, the exploitation of an optimal and/or robust design of the viscoelastic damped systems is a difficult task: laws of behaviour, complexity of FE models, etc, making very expensive any optimization procedure. In this work, the strategy proposed is a multiobjective optimization methodology combining metamodels: robust condensed models and artificial neural networks approach (ANNs).

Palavras-chave: multiobjective optimization; robust condensation; artificial neural networks; viscoelastic damping